A Systematic Literature Review of Weather-Driven Solar Energy Forecasting: Advanced Predictive Analytics

Abstract

The rapid adaptation of renewable energy resources such as solar energy to meet the growing global demand for energy necessitates the accurate and reliable forecasting of energy output to ensure grid stability and efficient resource allocation. In light of this, the topic was chosen to conduct a Systematic Literature Review (SLR) with the aim of gathering knowledge on predictive analytics methodologies, techniques, and approaches to weather-driven solar energy forecasting. The review objectives of ascertaining appropriate predictive analytics techniques significantly used variables, and best evaluation metrics for the same were met by the SLR conducted according to guidelines as proposed by B.Kitchenham [1]. The initial selection of 35 related studies was then narrowed to 15 for further detailed review. The results of the review indicate that Neural Networks (NN), Linear Regression, and Support Vector Machine models are the most used technologies, whereas the most significant variables considered in the studies were solar irradiation, temperature, historical power generation data, and relative humidity. The studies also emphasize the use of evaluation metrics such as RMSE and MAE for validating model accuracy. These findings provide valuable insights into predictive analytics in weather-driven solar energy forecasting and offer recommendations for best-suited approaches such as hybrid predictive models to implement in enhancing the accuracy and reliability of weather-driven solar energy forecasts.